Sponsorship or industry bias refers to the notion that industryfinancing of drug studies is strongly associated with findings thatfavour the industry or the sponsor’s product.1–3 Conflict ofinterest (COI), and specifically the financial type, is defined as‘a set of conditions in which professional judgment concerninga primary interest (such as a patient’s welfare or the validity ofresearch) tends to be unduly influenced by a secondary interest(such as financial gain)’.4 For the treatment of depression, industrybias has been clearly documented in comparisons between anti-depressants and placebo,5,6 where it seems to pivot aroundselective publication. However, the issues of detecting andquantifying the effects of sponsorship bias are more complicatedin head-to-head trials, looking at the comparative effectivenessor safety of two interventions. These trials are usually conductedafter a drug has been approved and consequently do not haveto be registered with a regulatory body. Hence, identifyingunpublished trials becomes infinitely more difficult, as it hingesentirely on the willingness of investigators to share their data. Thissignificant problem also implies that other methods have to beused for quantifying sponsorship bias than just scavenging forunpublished trials. Moreover, most head-to-head trials do notattempt to prove superiority of interventions, but non-inferiorityor equivalence and so even when they do find significantdifferences, these are usually small in magnitude. Consequently,we can expect the effects of sponsorship bias to be smaller andsubtler. For treatment for depression, some systematic reviewshave looked at the effects of sponsorship bias in head-to-headcomparisons between pharmacological treatments,7,8 but not forthe contrast between pharmacotherapy and psychotherapy. Recentmeta-analyses indicate these treatment options have comparableefficiency for depressive symptoms.9,10 Furthermore, a meta-analysis11 of direct comparisons between psychotherapy and placebopill found that the comparative effect size for psychotherapy wasin the range of what had been reported for antidepressants.6

Current treatment guidelines, such as those from the NationalInstitute for Health and Care Excellence (NICE),12 recommendboth for depression. Nevertheless, if the effects of sponsorship biasand financial COI are indeed that pervasive, we can conjecturethat they may also affect the contrast between a pharmacologicaland a psychological intervention. To the best of our knowledge,this question has not yet been considered for any mentaldisorder. Thus, we decided to examine the effects of sponsorshipbias and financial COI in randomised controlled trials (RCTs)directly comparing psychotherapy and medication in depression.Specifically, we predicted that each of these factors would beassociated with finding more positive effects for pharmacologicalthan psychological interventions.

Method

Identification and selection of studies

We used a database of papers on the psychological treatmentof depression described in detail elsewhere13 and that hasbeen used in a series of earlier published meta-analyses (www.evidencebasedpsychotherapies.org). This database has beencontinuously updated through comprehensive literature searches(covering studies published between 1966 and January 2015). Inthese searches, we examined abstracts from PubMed, PsycINFO,Embase and the Cochrane Register of Trials. These abstractswere identified by combining terms indicative of psychologicaltreatment and depression (both MeSH terms and text words, seeonline Supplement DS1). For this database, we also checked theprimary studies from earlier meta-analyses of psychologicaltreatments for depression to ensure that no published studies weremissed.

We included randomised trials in which the effects of apsychological treatment were directly compared with the effects

1

Sponsorship bias in the comparative efficacyof psychotherapy and pharmacotherapy for adultdepression: meta-analysisIoana A. Cristea, Claudio Gentili, Pietro Pietrini and Pim Cuijpers

BackgroundSponsorship bias has never been investigated for non-pharmacological treatments like psychotherapy.

AimsWe examined industry funding and author financial conflictof interest (COI) in randomised controlled trials directlycomparing psychotherapy and pharmacotherapy indepression.

MethodWe conducted a meta-analysis with subgroup comparisonsfor industry v. non-industry-funded trials, and respectively fortrial reports with author financial COI v. those without.

ResultsIn total, 45 studies were included. In most analyses,pharmacotherapy consistently showed significanteffectiveness over psychotherapy, g=70.11 (95% CI 70.21

to 70.02) in industry-funded trials. Differences betweenindustry and non-industry-funded trials were significant,a result only partly confirmed in sensitivity analyses.We identified five instances where authors of the originalarticle had not reported financial COI.

ConclusionsIndustry-funded trials for depression appear to subtly favourpharmacotherapy over psychotherapy. Disclosure of allfinancial ties with the pharmaceutical industry should beencouraged.

Declaration of interestP.P. received speaker fees from Jansen-Cilag.

Copyright and usageB The Royal College of Psychiatrists 2016.

The British Journal of Psychiatry

1–8. doi: 10.1192/bjp.bp.115.179275

Review article

of antidepressant medication in adults with a depressive disorder.Augmentation studies in which medication or psychotherapy werecombined and compared with one of them were excluded, as itwas impossible to disentangle the effects of psychotherapy andpharmacotherapy in these trials. Studies of in-patients were alsoexcluded. We further excluded maintenance or continuationstudies, aimed at people who had already recovered or partlyrecovered after an earlier treatment, again due to the lack of adirect comparison between psychotherapy and medication.Comorbid mental or somatic disorders were not used as exclusioncriteria, and neither was language.

Risk of bias assessment and data extraction

Trial risk of bias was assessed using four criteria of the CochraneCollaboration14 assessment tool: adequate generation of allocationsequence; concealment of allocation to conditions; the preventionof knowledge of the allocated intervention (masking of assessors);and dealing with incomplete outcome data (assessed as positivewhen intent-to-treat analyses, including all randomised patients,were conducted). We also computed a ‘risk of bias’ score for eachstudy, by giving one point to each criterion for which a studycould be rated as low risk of bias.

We extracted information about study financing from thearticle. Study financing was first coded into specific categories:government only; government and pharmaceutical industry;government and other sources; pharmaceutical industry only;pharmaceutical industry and other sources; other sources only;not reported. Industry support was defined as funding for the trialor provision of free medication, as in previous reviews.2 Othersources referred to non-governmental organisations or privatefoundations, which did not benefit from any industry support.This information was then recoded dichotomously into studieswith and without pharmaceutical industry funding.

Financial COI was defined as the receipt of financial supportor benefits of any type from the industry (employees, directcompensation, revenues from the industry, speaker fees, consultancy,serving on boards, etc). To extract this information, we first perusedwhat was reported in the article and extracted information aboutthe authors declaring receipt of benefits. For each study, thenumber of authors with financial COI was calculated. If no COIdisclosure was reported in the article, we proceeded to checkingother papers by the same author from the ones included in themeta-analysis. If there were none, or we still did not find anyinformation about financial COI in these either, we searchedPubMed for other publications by that author published 3 yearsbefore or no more than 2 years after the included paper. Thisperiod was chosen given that frequently there is a long intervalbetween the submission and acceptance of a paper. From thesepapers, we favoured those published in journals requiring authorsto declare COI (such as JAMA, BMJ), and in them checked whetherany of the authors of interest had reported financial COI.

We also coded additional aspects of the included studies, such asbaseline severity (computed as a weighted mean on the HamiltonRating Scale for Depression (HRSD) of the psychotherapy andpharmacotherapy arms), publication year, target group, type ofdepressive disorder, type of psychotherapy, class of medication.

Meta-analyses

We calculated and pooled the individual effect sizes with theprogram Comprehensive Meta-analysis (CMA; version 2.2.064),using a random-effects meta-analysis. For each comparisonbetween psychotherapy and pharmacotherapy, we calculated theeffect size indicating the difference between the two groups atpost-test (Hedges’ g). These were calculated by subtracting the

mean of the pharmacotherapy group from the mean of thepsychotherapy group, dividing the result by the pooled standarddeviation, and correcting for small sample bias.15 If a studyincluded more than one comparison between a type ofpsychotherapy and a drug, these were averaged at the study level.Sensitivity analyses were also conducted using only thecomparisons with effect size most favourable to psychotherapyand pharmacotherapy, respectively. We only used instruments thatexplicitly measured depressive symptoms, such as HRSD or theBeck Depression Inventory (BDI). If means and standarddeviations were not reported and could not be obtained fromthe authors, we used the procedures recommended by Borensteinet al16 to transform dichotomous data into the standardisedmean difference or used other statistics, such as t-values or exactP-values to calculate the standardised mean difference. Wecalculated the I 2-statistic as an indicator of heterogeneity inpercentages. A value of 0% indicates no observed heterogeneity,whereas larger values indicate increasing heterogeneity, with25% as low, 50% as moderate and 75% as high heterogeneity.17

We calculated 95% confidence intervals around I 2,18 using thenon-central w2-based approach with the heterogi module forStata.19 We tested for publication bias by inspecting the funnelplot on primary outcome measures and by Duval & Tweedie’s trimand fill procedure,20 which yields an estimate of the effect size afterthe publication bias has been taken into account. We also conductedEgger’s test of the intercept to quantify the bias captured by thefunnel plot and to test whether it was significant. Outliers weredefined as effect sizes for which the 95% confidence interval wasoutside the 95% confidence interval of the pooled studies.

We used a mixed-effects model to test whether the effect sizesof the studies with industry funding differed from those of thestudies without it, and whether studies authored by individualswith financial COI differed from those where this was notreported. In this model, studies within subgroups are pooled usingthe random-effects model, but tests for significant differencesbetween subgroups are conducted using a fixed-effects model.21

We conducted a series of sensitivity analyses to examine therobustness of our findings. In them, we looked at the followingsubgroups of studies: studies with low risk of bias on three or fourcriteria; published from 2000 onwards (information aboutfinancial COI is more rare and harder to find for older studies);with a selective serotonin reuptake inhibitor (SSRI) aspharmacotherapy; with cognitive–behavioural therapy (CBT) aspsychotherapy; conducted on patients with a diagnosis of majordepressive disorder (MDD, excluding patients with dysthymia orother diagnosis); on adults only (excluding specific target groups).

Results

Selection and inclusion of studies

After examining a total of 17 061 abstracts (12 196 after removalof duplicates), we retrieved 1756 full-text papers for furtherconsideration. Figure 1 presents a flow chart describing theinclusion process. We excluded 1711 of the retrieved papers(reasons given in the flow chart). A total of 45 studies,22–66

9 of which included two comparisons between a form ofpsychotherapy and medication, met the inclusion criteria.

Characteristics of the included studies

Selected characteristics of the 45 included studies are presented inonline Table DS1. Most studies were targeted at adults in general(n= 37), with a diagnosis of MDD (n= 37), and used CBT (n= 24comparisons) or interpersonal therapy (n= 12 comparisons) aspsychotherapy, and SSRIs (n= 20) or tricyclics (n= 12) as

2

Cristea et al

Sponsorship bias meta-analysis

medication. Six studies had low risk of bias on all four criteriaconsidered, 9 studies on three criteria, 25 studies had low risk ofbias on one or two criteria, and 5 studies on none.

Regarding study funding, 11 studies reported fundingexclusively from the government, 13 from both the governmentand pharmaceutical industry, 6 from both government andother sources (such as philanthropic foundations, universities,non-governmental organisations), 4 exclusively from the industry,3 were financed by the industry and other sources and 4 by othersources exclusively. Four studies did not report any funding.Recoding study financing dichotomously resulted into 20 studies(44%) with pharmaceutical industry support (for 11 of thesesupport involved just providing free medication) and 21 (47%)without (the 4 studies with no funding reported were notconsidered).

For financial COI, ten studies included authors who reporteda financial COI. We additionally uncovered five other studies. Thepercentage of authors with financial COI of the total number ofauthors in each study ranged from 8 to 92%. For the remaining30 studies no financial COI was reported and we did not findany further information.

Overall effect size

The overall effect size for the comparison between psychotherapyand pharmacotherapy was non-significant, g=70.01 (95% CI70.12 to 0.09), I 2 = 61% (95% CI 42 to 71), in line with previousmeta-analyses.9,67 Duval & Tweedie’s trim and fill procedureimputed four missing studies, leading to a non-significantg=70.09, 95% CI 70.20 to 0.02. However, Egger’s test did notindicate any significant asymmetry of the funnel plot (P= 0.19).

Industry funding

For the 20 studies with industry funding, pharmacotherapy wassignificantly more effective than psychotherapy, g=70.11 (95%

CI 70.21 to 70.02), with low heterogeneity (I 2 = 19%) (Table1, Fig. 2). For the 21 trials with no industry support, there wereno differences between pharmacotherapy and psychotherapy,g= 0.10 (95% CI 70.09 to 0.29), and heterogeneity was moderate(I 2 = 73%). The difference between studies with and withoutindustry funding was significant (P= 0.05). In the 11 trials whereindustry support was solely free medication, the differencebetween psychotherapy and pharmacotherapy was not significant,g=70.07 (95% CI 70.20 to 0.05). With their exclusion, thepattern of results remained the same and differences betweenindustry and non-industry-funded studies were significant(P= 0.028). In analysis with outliers excluded, or including onlyone comparison per study, pharmacotherapy was more effectivethan psychotherapy in industry-funded trials, but the differencebetween these and non-industry-funded ones was significant onlywhen the analyses included the comparison most favourable topsychotherapy.

There was no indication of publication bias (online Fig. DS1)for the studies with industry funding, neither with Duval &Tweedie’s trim and fill procedure, nor with Egger’s test. For thestudies without industry funding, Duval & Tweedie’s trim and fillprocedure imputed four studies, leading to a non-significantg=70.06 (95% CI 70.27 to 0.15), but Egger’s test did notindicate any significant asymmetry of the funnel plot.

Sensitivity analyses

These analyses partially replicated the pattern found in the mainanalyses (Table 1). For the studies with industry support, therewas a small but significant difference between psychotherapyand pharmacotherapy favouring the latter for studies with lowrisk of bias on three or four of the criteria considered, studiespublished from 2000 onwards, using SSRIs as medication, oraimed at adults in general (effect sizes ranged from 70.16 to70.07). There were no differences for studies employing CBT orfor participants with MDD. For the trials with no industrysupport, there were no differences between psychotherapy andpharmacotherapy in any of the analyses. Differences betweenindustry and non-industry-funded studies were statisticallysignificant for studies aimed at adults in general (P= 0.033) andfor trials using SSRIs (P= 0.036).

Author financial COI

For the 15 trials where we identified at least one author as havingreceived financial benefits from the pharmaceutical industry, thedifference between pharmacotherapy and psychotherapy, favouringpharmacotherapy, closely approached statistical significance,g=70.13 (95% CI 70.27 to 0.003, P= 0.054), with moderateheterogeneity (I 2 = 51%) (Table 2, Fig. 3). In the 30 trials whereno financial COI was reported and we were unable to find anyinformation about it, there were no significant differences betweenthe two treatments, g= 0.05 (95% CI 70.08 to 0.19), and hetero-geneity was moderate (I 2 = 60%). The difference between trialswhere author financial COIs were documented and those wherethey were not was close to statistical significance (P= 0.057).

Exclusion of outliers rendered differences within the trialswith authors financial COI, and between these and trials wherethere was no reported COI not significant. Analyses includingonly the comparison favouring psychotherapy in studies withtwo comparisons showed significant differences between trialswith author COI and those with no information on this(P= 0.037). When analyses were restricted to self-report measures,there was a more sizable significant advantage of pharmacotherapyover psychotherapy, g=70.26 (95% CI 70.42 to 70.11) within

3

17 061 references identifiedby literature search:– Pub Med: 4007– PsycINFO 3147– Embase: 5912– Cochrane: 3995

After removal of duplicates:12 196 abstracts

1756 publications retrieved

45 randomised trialscomparing psychotherapywith pharmacotherapyfor adult depression

10 440 excluded based ontitle and abstract

Excluded: 1711– Studies with adolescents (74)– No random assignment (62)– Not only depression (222)– No psychotherapy (200)– No comparison condition (135)– Maintenance trial (113)– ES cannot be calculated (23)– Secondary papers on same

study (323)– Trial but wrong comparison (395)– Study protocol (43)– Other reason (121)

6

7

6

7

6

Fig. 1 Flow chart of selection and inclusion process, followingthe PRISMA statement.

Cristea et al

4

Table 1 Effects of studies comparing psychotherapy and pharmacotherapy for adult depression: industry funding

With industry funding for the study Without industry funding for the study

Ncomp ga (95% CI) I 2 (95% CI) Ncomp ga (95% CI) I 2(95% CI) Pb

All studies 20 70.11 (70.21 to 70.02) 19 (0 to 52) 21 0.10 (70.09 to 0.29) 73 (55 to 81) 0.05

Outliers removedc 20 70.11 (70.21 to 70.02) 19 (0 to 52) 17 70.003 (70.13 to 0.13) 31 (0 to 61) 0.186

One effect size/studyd

(most favourable to psychotherapy) 20 70.11 (70.20 to 70.01) 18 (0 to 52) 21 0.14 (70.06 to 0.33) 74 (58 to 82) 0.026

One effect size/studyd

(most favourable to pharmacotherapy) 20 70.12 (70.21 to 70.02) 21 (0 to 54) 21 0.06 (70.13 to 0.25) 73 (56 to 81) 0.099

Studies with free medication excluded 9 70.17 (70.33 to 70.02) 33 (0 to 68) 21 0.10 (70.09 to 0.29) 73 (55 to 81) 0.028

Only clinician-based measures 18 70.09 (70.21 to 0.03) 37 (0 to 63) 16 0.12 (70.07 to 0.31) 62 (24 to 76) 0.063

Only self-report measures 12 70.10 (70.27 to 0.06) 37 (0 to 67) 17 0.09 (70.15 to 0.32) 77 (62 to 85) 0.202

Sensitivity analyses

Only high-quality studies 10 70.14 (70.26 to 70.02) 14 (0 to 59) 5 0.12 (70.36 to 0.59) 89 (77 to 94) 0.308

Studies published from 2000 onwards 17 70.13 (70.23 to 70.03) 17 (0 to 54) 14 0.04 (70.21 to 0.29) 80 (65 to 86) 0.208

Only selective serotonin reuptake inhibitors 8 70.16 (70.27 to 70.04) 0 (0 to 56) 11 0.16 (70.11 to 0.43) 75 (49 to 85) 0.036Only cognitive–behavioural therapy 8 70.07 (70.27 to 0.14) 35 (0 to 70) 14 0.12 (70.10 to 0.34) 66 (31 to 79) 0.219

Only major depressive disorder 14 70.10 (70.22 to 0.02) 12 (0 to 53) 20 0.11 (70.09 to 0.31) 74 (57 to 82) 0.085

Studies aimed at adults in general 16 70.12 (70.23 to 70.02) 19 (0 to 55) 18 0.10 (70.08 to 0.28) 64 (32 to 77) 0.033

Ncomp, number of comparisons.a. According to the random-effects model. A positive effect indicates superiority of psychotherapy. Significant values are in bold.b. The P-values in this column indicate whether the difference between the effect sizes in the group of studies with industry funding differs from those without industry funding.Significant values are in bold.c. Outliers were defined as studies in which the 95% confidence interval was outside the 95% confidence interval of the pooled studies. Above the 95% confidence interval (favouringpsychotherapy) was Faramarzi et al,35 Moradveisi et al,50 Rush et al.56 Below the 95% CI (favouring pharmacotherapy): Sharp et al.61

d. Studies with more than one comparison: David et al,28 Dimidjian et al31 Elkin et al34 Markowitz et al,43 McLean et al,46 Mohr et al,48 Mynor-Wallis et al,52 Quilty et al55 and Scott& Freeman.59

Statistics for each studyGroup by Hedges’ Lower UpperPharma support Study g Limit Limit P

Bedi et al (2000)24

Blackburn & Moore (1997)25

David et al (2008)28

Dunlop et al (2012)32

Elkin et al (1989)34

Faramarzi et al (2008)35

Hegerl et al (2010)38

Hollon et al (1992)39

McLean & Hakstian (1979)46

Menchetti et al (2014)47

Mohr et al (2001)49

Moradveisi et al (2013)50

Mynors-Wallis et al (2000)52

Quilty et al (2008)55

Rush et al (1977)56

Salminen et al (2008)57

Scott & Freeman (1992)59

Shamsaei et al (2008)60

Sharp et al (2010)61

Weissman et al (1979)64

Zu et al (2014)66

Barber et al (2012)22

Barrett et al (2001)23

Blom et al (2007)26

Browne et al (2002)27

Dekker (2008)29

DeRubeis et al (2005)30

Dimidjian et al (2006)31

Finkenzeller et al (2009)36

Frank et al (2011)37

Jarrett et al (1999)40

Keller et al (2000)41

Kennedy et al (2007)42

Markowitz et al (2005)43

Martin et al (2001)44

Miranda et al (2003)48

Murphy et al (1984)51

Mynors-Wallis et al (1995)53

Parker et al (2013)54

Thompson et al (2001)63

Williams et al (2000)65

NoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesOverall

70.040.090.04

70.2170.08

0.9470.04

0.080.330.30

70.101.05

71.1870.50

0.9070.11

0.3070.4770.48

0.240.600.10

70.1170.12

0.1370.2570.4070.1570.16

0.1770.1070.2270.0470.5170.6870.6470.24

0.320.210.580.31

70.1570.1170.07

70.4670.4770.3270.7070.44

0.4170.4370.3970.02

0.0470.76

0.5970.6370.93

0.2170.6570.2270.9170.7570.6370.0870.0970.5870.4370.3570.5370.7970.4870.6070.3770.3470.6770.2371.3071.2571.3870.5370.2470.3170.1770.1770.3970.2170.15

0.390.650.410.270.281.470.360.550.690.560.561.520.27

70.081.590.430.83

70.0370.22

1.101.290.290.370.190.620.04

70.000.190.280.720.150.240.150.27

70.110.100.050.880.731.320.800.08

70.020.01

0.870750.820.390.660.000.850.730.060.020.770.000.430.020.010.690.260.040.000.590.090.290.670.450.590.090.050.400.470.530.440.360.680.200.020.090.110.270.420.130.210.200.020.10

72.00 71.00 0.00 1.00 2.00

Favours pharmacotherapy Favours psychotherapy

Hedges’ g and 95% CI

Fig. 2 Standardised effect sizes of comparisons between psychotherapy and pharmacotherapy for adult depression, with and withoutindustry funding.

Sponsorship bias meta-analysis

the studies with financial COI, and differences between the twosubgroups were significant (P= 0.002). We also examined COIseparately when financial support was given to the first or lastauthor. In studies where the first author had financial COI, thedifference between psychotherapy and pharmacotherapy,favouring the latter, was near statistical significance (P= 0.056).There were no significant differences for last author financial COI.

There was no indication of publication bias (online Fig. DS2)for the studies whose authors had financial COI, neither withDuval & Tweedie’s trim and fill procedure, nor with Egger’s test.For the studies with no information about financial COI, theDuval & Tweedie trim and fill procedure imputed 9 studies, leadingto a non-significant g=70.13 (95% CI 70.28 to 0.02) and Egger’stest indicated an asymmetrical funnel plot (P= 0.018).

Sensitivity analyses

We did not replicate the pattern found in the main analyses,except for the trials published after 2000, where studies withauthor financial COI showed a difference favouring pharmaco-therapy over psychotherapy that was close to statistical significance(P= 0.054). Differences between the trials with author COI andwithout were borderline significant for studies on patients withMDD (P= 0.061).

Discussion

Summary of main findings

In this meta-analysis, we included published reports of RCTsdirectly comparing a psychological and a pharmacological treatmentfor the acute treatment of depression and examined the potentialeffects of sponsorship bias and author financial COI on thesecomparisons. We focused on two types of analyses: whether onetreatment option was more effective than the other within thesubgroup of studies with industry support or where authors had

financial COI, and whether the estimations of the comparativeeffectiveness of these treatments differed between studies withindustry support v. studies without, and studies where authorshad financial COI v. studies for which we did not have informationon this. Our results showed that in studies with industry supportpharmacological treatments were more effective than psychologicalones, whereas differences between the two were non-significantfor studies with no industry support. Moreover, there was asignificant difference between industry-funded v. non-industry-funded studies. Additional analyses excluding outliers, excludingstudies where the financial support was solely as free medication,and most sensitivity analyses looking at specific subgroups ofstudies confirmed the pattern of industry-funded studies findingpharmacotherapy more effective than psychotherapy. We alsonoted that heterogeneity estimates in almost all analyses on theindustry-funded studies were small, which increases confidencein the robustness of the effect size estimation. However,differences between the estimates of industry sponsored andnon-industry sponsored trials were only significant in some ofthe analyses and consequently the clinical relevance of thedifference we found in the industry-sponsored subgroup couldbe limited and most likely does not reach the standard for aclinically relevant effect.68

This was to be expected since both pharmacotherapy andpsychotherapy were consistently shown to be effective in thetreatment of depression9,69,70 so it would have been disconcertingto find something other than subtle, small magnitude differences.Still, our results corroborate previous reports71 about thedeleterious effects of sponsorship bias on treatment outcomeresearch, as they show a potential additional trend of industry-funded research to favour pharmacotherapy over psychologicaltreatments. It is worth noting that this pattern was present evenif most studies benefited from only partial industry sponsorship.

Conversely, author financial COI, involving authors’ personal,economic ties with the industry, may not be connected to thefinancing of the trial at all, hence its relationship to outcomes

5

Table 2 Effects of studies comparing psychotherapy and pharmacotherapy for adult depression: financial conflict of interest (COI)

Studies with authors with financial COI Studies with no information on financial COI

Ncomp ga (95% CI) P I 2 (95% CI) Ncomp ga (95% CI) P I 2 (95% CI) Pb

All studies 15 70.13 (70.27 to 0.003) 0.054 51 (0 to 71) 30 0.06 (70.08 to 0.19) 60 (36 to 72) 0.057

Outliers removedc 14 70.09 (70.21 to 0.04) 35 (0 to 65) 27 70.06 (70.16 to 0.03) 17 (0 to 48) 0.773

One effect size/studyd

(most favourable to psychotherapy) 15 70.13 (70.26 to 0.009) 50 (0 to 71) 30 0.08 (70.06 to 0.22) 62 (40 to 74) 0.037

One effect size/studyd

(most favourable to pharmacotherapy) 15 70.14 (70.28 to –0.003) 52 (0 to 72) 30 0.03 (70.11 to 0.16) 60 (36 to 72) 0.087

First author financial COI 9 70.17 (70.33 to 0.005) 0.056 43 (0 to 72) 36 0.02 (70.10 to 0.13) 59 (37 to 71) 0.080

Last author financial COI 5 70.01 (70.24 to 0.22) 49 (0 to 80) 40 70.02 (70.13 to 0.09) 59 (38 to 70) 0.949

Only clinician-based measures 12 70.06 (70.23 to 0.10) 48 (0 to 72) 25 0.03 (70.11 to 0.17) 57 (26 to 71) 0.391

Only self-report measures 9 70.26 (70.42 to 70.11) 23 (0 to 64) 23 0.12 (70.06 to 0.29) 65 (42 to 77) 0.002

Sensitivity analyses

Only low risk of bias studies 9 70.10 (70.30 to 0.10) 68 (19 to 82) 6 0.04 (70.33 to 0.41) 81 (51 to 90) 0.515

Studies published from 2000 onward 15 70.13 (70.27 to 0.003) 51 (0 to 71) 16 0.03 (70.18 to 0.24) 74 (53 to 83) 0.199

Only selective serotonin reuptake inhibitors 10 70.09 (70.24 to 0.06) 41 (0 to 71) 11 0.11 (70.18 to 0.41) 77 (55 to 86) 0.233

Only cognitive–behavioural therapy 5 70.11 (70.38 to 0.15) 20 (0 to 71) 19 0.07 (70.11 to 0.25) 62 (30 to 76) 0.254

Only major depressive disorder 9 70.17 (70.39 to 0.05) 61 (0 to 79) 28 0.08 (70.06 to 0.22) 61 (36 to 73) 0.061

Studies aimed at adults in general 13 70.08 (70.22 to 0.06) 38 (0 to 67) 24 0.02 (70.12 to 0.17) 58 (28 to 73) 0.322

Ncomp, number of comparisons.a. According to the random-effects model. A positive effect indicates superiority of psychotherapy. Significant values are in bold.b. The P-values in this column indicate whether the difference between the effect sizes in the group of studies with author financial COI differ from those where we did not haveinformation about author COI.Significant values are in bold.c. Outliers were defined as studies in which the 95% confidence interval was outside the 95% confidence interval of the pooled studies. Above the 95% confidence interval(favouring psychotherapy) was Faramarzi et al,35 Moradveisi et al,50 Rush et al.56 Below the 95% CI (favouring pharmacotherapy): Sharp et al.61

d. Studies with more than one comparison: David et al,28 Dimidjian et al31 Elkin et al34 Markowitz et al,43 McLean et al,46 Mohr et al,48 Mynor-Wallis et al,52 Quilty et al55 and Scott& Freeman.59

Cristea et al

could be more complex. The most noteworthy result is that weidentified five instances where one or more of the authors of theoriginal article had a financial COI and had not reported it. Wewere able to verify this information by looking at other publishedtrials involving the same authors in the same period. Studies inwhich one or more of the authors had financial ties with theindustry resulted in a small and close to statistical significanceadvantage of pharmacotherapy over psychotherapy. Differencesbetween this subgroup of trials and those for which we did nothave information about author financial COI also closelyapproached statistical significance. This pattern was preserved insome of the additional sensitivity analyses (most notably whenthe financial COI was connected to the first author and whenwe restricted outcomes to self-report measures), but not inothers. Clearly, the difference is small and unstable, so it is unlikelyto be clinically relevant. We underscore that we considered anytype of industry compensation, regardless of the particularpharmaceutical company and of the type of compensation (e.g.employment with the company, author grants, speaker fees).

ImplicationsSeveral possible mechanisms have been proposed2,8,71 and it islikely that a constellation of factors could explain these effects,both for industry support and for author financial COI.

Our results certainly do not establish a clear implication, ofsolid clinical significance, that the presence of industry supportor that of authors with financial COI are responsible for morefavourable outcomes for an industry option (pharmacotherapy)over a non-industry one (psychotherapy). Nonetheless, they doraise doubt that there might be such bias at play, thus adding toan ever-growing literature painfully pointing to the pervasivenessof industry influences on treatment outcome research. Also, andeven more alarmingly, for financial COI, our results point to thefact that the necessary information for evaluating such a biascould be missing from a non-negligible portion of published trials.A possible remedy could involve journal editorial boards takingtransparency one step further to full disclosure and asking authorsof trials involving medication as one of the treatments to declareany and indeed all financial ties with the pharmaceutical industry,

6

Statistics for each studyGroup by Hedges’ Lower UpperAuthor COI Study g Limit Limit P

Hedges’ g and 95% CI

NoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoNoYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesOverall

Bedi et al (2000)24

Blackburn & Moore (1997)25

Blom et al (2007)26

Browne et al (2002)27

David et al (2008)28

Dekker (2008)29

DeRubeis et al (2005)30

Dunner et al (1996)33

Elkin et al (1989)34

Faramarzi et al (2008)35

Finkenzeller et al (2009)36

Hollon et al (1992)39

Jarrett et al (1999)40

McKnight et al (1992)45

McLean & Hakstian (1979)46

Miranda et al (2003)48

Mohr et al (2001)49

Moradveisi et al (2013)50

Murphy et al (1984)51

Mynors-Wallis et al (2000)52

Mynors-Wallis et al (1995)53

Quilty et al (2008)55

Rush et al (1977)56

Schulberg et al (1996)58

Scott & Freeman (1992)59

Shamsaei et al (2008)60

Sloane et al (1985)62

Thompson et al (2001)63

Weissman et al (1979)64

Zu et al (2014)66

Barber et al (2012)22

Barrett et al (2001)23

Dimidjian et al (2006)31

Dunlop et al (2012)32

Frank et al (2011)37

Hegerl et al (2010)38

Keller et al (2000)41

Kennedy et al (2007)42

Markowitz et al (2005)43

Martin et al (2001)44

Menchetti et al (2014)47

Parker et al (2013)54

Salminen et al (2008)57

Sharp et al (2010)61

Williams et al (2000)65

70.040.090.13

70.250.04

70.4070.1570.3470.08

0.940.170.08

70.2270.34

0.3370.2470.10

1.050.320.21

70.1870.50

0.9070.03

0.3070.47

0.140.310.240.600.06

70.1170.1370.1670.2170.1070.0470.0470.5170.6870.64

0.300.58

70.1170.4870.1570.1370.04

70.4670.4770.3570.5370.3170.7970.4871.1870.44

0.3170.3770.3970.6770.9470.0270.5370.76

0.5970.2470.3170.6370.93

0.2170.3270.2270.9170.6070.1770.6370.0870.0870.5870.4370.6070.7070.3470.4370.2371.3071.2571.38

0.0470.1770.6570.7570.3970.2770.13

0.390.650.620.040.41

70.000.190.500.281.470.720.550.240.250.690.050.561.520.880.730.27

70.081.590.260.83

70.030.890.801.101.290.190.370.190.280.270.150.360.150.27

70.110.100.561.320.43

70.220.080.000.06

0.870.750.590.090.820.050.400.420.660.000.530.730.360.250.060.110.770.000.270.420.430.020.010.820.260.040.700.210.590.090.410.670.450.470.390.440.850.680.200.020.090.020.130.690.000.200.050.44

72.00 71.00 0.00 1.00 2.00

Favours pharmacotherapy Favours psychotherapy

Fig. 3 Standardised effect sizes of comparisons between psychotherapy and pharmacotherapy for adult depression, with and withoutauthor financial conflict of interest (COI).

Sponsorship bias meta-analysis

regardless of whether they deem these relevant or not to the trialor paper in question. The policy seems to be already effectivelyenforced by some flagship journals such as the American Journalof Psychiatry and JAMA. This measure would allow a reliableassessment of potentially biasing effects of author financialsupport from the industry.

Limitations

There are a number of limitations to this meta-analysis. Asexpected, most analyses were affected by a moderate or highdegree of heterogeneity, particularly for studies without industryfunding or for those without information about author financialCOI. Moreover, while few, the outliers identified provided estimationsvery far from the pooled effect size. Higher heterogeneity and thepresence of most outliers in the no industry support/noinformation about financial COI trials could have acted as aconfounding factor for between-subgroup comparisons,particularly since the differences we found were small. We onlylooked at published trials and did not attempt to identify‘abandoned’ trials from trial registries or from investigators. Wealso did not distrust what investigators declared as study funding.A significant portion of the included studies had high or uncertainrisk of bias but, interestingly, trials with low risk of bias were stillmore likely to find pharmacotherapy to be more effective thanpsychotherapy.

We cannot exclude that there might be additional instances ofundisclosed financial COI that we did not uncover that could havehad an impact on the pattern of results, since an exhaustive searchof all papers published by all authors was impossible to ensureand, particularly for older papers, our search options were morelimited. Finally, we only looked at financial COI related to thepharmaceutical industry, but similar concerns have been recentlyraised about psychotherapy. We did not examine financial COIrelated to psychotherapy, such as royalties from treatmentmanuals, benefits from psychotherapy training, courses or work-shops. Whereas assessing financial COI from the pharmaceuticalindustry is rather straightforward and several previous reviewshave examined it, there are no proposed or established tools orguidelines for assessing direct financial payback from psycho-therapy. As such, not only is there no consensus as to what shouldbe tallied as financial COI related to psychotherapy, but thisinformation is missing in most published articles,72 rendering itspossible estimates uncertain.

Ioana A. Cristea, Department of Clinical Psychology and Psychotherapy, BabesBolyai University, Cluj-Napoca, Romania, and Department of General Psychology,University of Padova, Padova, Italy; Claudio Gentili, Department of GeneralPsychology, University of Padova, Padova, Italy, Pietro Pietrini, IMT Institute forAdvanced Studies, Lucca, Italy; Pim Cuijpers, Department of Clinical Psychology, VUUniversity and EMGO Institute for Health and Care Research, Amsterdam, TheNetherlands

Correspondence: Ioana A. Cristea, Babes-Bolyai University, Republicii 37Street, Cluj-Napoca, Romania 400015. Email: ioana.cristea@ubbcluj.ro

First received 10 Dec 2015, final revision 6 Mar 2016, accepted 25 Jun 2016

References

1 Bekelman JE, Li Y, Gross CP. Scope and impact of financial conflicts ofinterest in biomedical research: a systematic review. JAMA 2003; 289:454–65.

2 Lundh A, Sismondo S, Lexchin J, Busuioc OA, Bero L. Industry sponsorshipand research outcome. Cochrane Database Syst Rev 2012; 12: MR000033.

3 Lexchin J, Bero LA, Djulbegovic B, Clark O. Pharmaceutical industrysponsorship and research outcome and quality: systematic review. BMJ2003; 326: 1167–70.

4 Thompson DF. Understanding financial conflicts of interest. N Engl J Med1993; 329: 573–6.

5 Melander H, Ahlqvist-Rastad J, Meijer G, Beermann B. Evidence b(i)asedmedicine–selective reporting from studies sponsored by pharmaceuticalindustry: review of studies in new drug applications. BMJ 2003; 326: 1171–3.

6 Turner EH, Matthews AM, Linardatos E, Tell RA, Rosenthal R. Selectivepublication of antidepressant trials and its influence on apparent efficacy.N Engl J Med 2008; 358: 252–60.

7 Gartlehner G, Morgan L, Thieda P, Fleg A. The effect of study sponsorship ona systematically evaluated body of evidence of head-to-head trials wasmodest: secondary analysis of a systematic review. J Clin Epidemiol 2010;63: 117–25.

8 Baker CB, Johnsrud MT, Crismon ML, Rosenheck RA, Woods SW. Quantitativeanalysis of sponsorship bias in economic studies of antidepressants. Br JPsychiatry 2003; 183: 498–506.

9 Cuijpers P, Sijbrandij M, Koole SL, Andersson G, Beekman AT, Reynolds CF.The efficacy of psychotherapy and pharmacotherapy in treating depressiveand anxiety disorders: a meta-analysis of direct comparisons. WorldPsychiatry 2013; 12: 137–48.

10 Imel ZE, Malterer MB, McKay KM, Wampold BE. A meta-analysis ofpsychotherapy and medication in unipolar depression and dysthymia.J Affect Disord 2008; 110: 197–206.

11 Cuijpers P, Turner EH, Mohr DC, Hofmann SG, Andersson G, Berking M, et al.Comparison of psychotherapies for adult depression to pill placebo controlgroups: a meta-analysis. Psychol Med 2014; 44: 685–95.

12 National Institute for Health and Clinical Excellence. Treatment andManagement of Depression in Adults, Including Adults with a ChronicPhysical Health Problem. NICE, 2009.

13 Cuijpers P, van Straten A, Warmerdam L, Andersson G. Psychologicaltreatment of depression: a meta-analytic database of randomized studies.BMC Psychiatry 2008; 8: 36.

14 Higgins JPT, Altman DG, Gotzsche PC, Juni P, Moher D, Oxman AD, et al.The Cochrane Collaboration’s tool for assessing risk of bias in randomisedtrials. BMJ 2011; 343: d5928.

15 Hedges LV, Olkin I. Statistical Methods for Meta-Analysis. Academic Press,1985.

16 Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Introduction toMeta-Analysis. Wiley, 2009.

17 Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistencyin meta-analyses. BMJ 2003; 327: 557–60.

18 Ioannidis JPA, Patsopoulos NA, Evangelou E. Uncertainty in heterogeneityestimates in meta-analyses. BMJ 2007; 335: 914–6.

19 Orsini N, Bottai M, Higgins J, Buchan I. HETEROGI: Stata Module to QuantifyHeterogeneity in a Meta-Analysis. Boston College Department of Economics,2006 (http://econpapers.repec.org/software/bocbocode/s449201.htm).

20 Duval S, Tweedie R. Trim and fill: a simple funnel-plot-based method oftesting and adjusting for publication bias in meta-analysis. Biometrics 2000;56: 455–63.

21 Borenstein M, Hedges LV, Higgins JPT, Rothstein HR. Introduction toMeta-Analysis. Wiley, 2009.

22 Barber JP, Barrett MS, Gallop R, Rynn MA, Rickels K. Short-term dynamicpsychotherapy versus pharmacotherapy for major depressive disorder: arandomized, placebo-controlled trial. J Clin Psychiatry 2012; 73: 66–73.

23 Barrett JE, Williams JW, Oxman TE, Frank E, Katon W, Sullivan M, et al.Treatment of dysthymia and minor depression in primary care: a randomizedtrial in patients aged 18 to 59 years. J Fam Pract 2001; 50: 405–12.

24 Bedi N, Chilvers C, Churchill R, Dewey M, Duggan C, Fielding K, et al.Assessing effectiveness of treatment of depression in primary care. Partiallyrandomised preference trial. Br J Psychiatry 2000; 177: 312–8.

25 Blackburn IM, Moore RG. Controlled acute and follow-up trial of cognitivetherapy and pharmacotherapy in out-patients with recurrent depression.Br J Psychiatry 1997; 171: 328–34.

26 Blom MBJ, Jonker K, Dusseldorp E, Spinhoven P, Hoencamp E, Haffmans J,et al. Combination treatment for acute depression is superior only whenpsychotherapy is added to medication. Psychother Psychosom 2007; 76:289–97.

27 Browne G, Steiner M, Roberts J, Gafni A, Byrne C, Dunn E, et al. Sertralineand/or interpersonal psychotherapy for patients with dysthymic disorder inprimary care: 6-month comparison with longitudinal 2-year follow-up ofeffectiveness and costs. J Affect Disord 2002; 68: 317–30.

28 David D, Szentagotai A, Lupu V, Cosman D. Rational emotive behaviortherapy, cognitive therapy, and medication in the treatment of majordepressive disorder: a randomized clinical trial, posttreatment outcomes,and six-month follow-up. J Clin Psychol 2008; 64: 728–46.

7

Cristea et al

29 Dekker JJM, Koelen JA, Van HL, Schoevers RA, Peen J, Hendriksen M, et al.Speed of action: the relative efficacy of short psychodynamic supportivepsychotherapy and pharmacotherapy in the first 8 weeks of a treatmentalgorithm for depression. J Affect Disord 2008; 109: 183–8.

30 DeRubeis RJ, Hollon SD, Amsterdam JD, Shelton RC, Young PR, Salomon RM,et al. Cognitive therapy vs medications in the treatment of moderate tosevere depression. Arch Gen Psychiatry 2005; 62: 409–16.

31 Dimidjian S, Hollon SD, Dobson KS, Schmaling KB, Kohlenberg RJ, Addis ME,et al. Randomized trial of behavioral activation, cognitive therapy, andantidepressant medication in the acute treatment of adults with majordepression. J Consult Clin Psychol 2006; 74: 658–70.

32 Dunlop BW, Kelley ME, Mletzko TC, Velasquez CM, Craighead WE, MaybergHS. Depression beliefs, treatment preference, and outcomes in a randomizedtrial for major depressive disorder. J Psychiatr Res 2012; 46: 375–81.

33 Dunner DL, Schmaling KB, Hendrickson H, Becker J, Lehman A, Bea C.Cognitive therapy versus fluoxetine in the treatment of dysthymic disorder.Depression 1996; 4: 34–41.

34 Elkin I, Shea MT, Watkins JT, Imber SD, Sotsky SM, Collins JF, et al. NationalInstitute of Mental Health Treatment of Depression Collaborative ResearchProgram. General effectiveness of treatments. Arch Gen Psychiatry 1989; 46:971–82; discussion 983.

35 Faramarzi M, Alipor A, Esmaelzadeh S, Kheirkhah F, Poladi K, Pash H.Treatment of depression and anxiety in infertile women: cognitive behavioraltherapy versus fluoxetine. J Affect Disord 2008; 108: 159–64.

36 Finkenzeller DW, Zobel I, Rietz S, Schramm E, Berger M. InterpersonellePsychotherapie und Pharmakotherapie bei Post-Stroke-Depression.Nervenarzt 2009; 80: 805–12.

37 Frank E, Cassano GB, Rucci P, Thompson WK, Kraemer HC, Fagiolini A, et al.Predictors and moderators of time to remission of major depression withinterpersonal psychotherapy and SSRI pharmacotherapy. Psychol Med 2011;41: 151–62.

38 Hegerl U, Hautzinger M, Mergl R, Kohnen R, Schutze M, Scheunemann W,et al. Effects of pharmacotherapy and psychotherapy in depressedprimary-care patients: a randomized, controlled trial including a patients’choice arm. Int J Neuropsychopharmacol 2010; 13: 31–44.

39 Hollon SD, DeRubeis RJ, Evans MD, Wiemer MJ, Garvey MJ, Grove WM, et al.Cognitive therapy and pharmacotherapy for depression. Singly and incombination. Arch Gen Psychiatry 1992; 49: 774–81.

40 Jarrett RB, Schaffer M, McIntire D, Witt-Browder A, Kraft D, Risser RC.Treatment of atypical depression with cognitive therapy or phenelzine: adouble-blind, placebo-controlled trial. Arch Gen Psychiatry 1999; 56: 431–7.

41 Keller MB, McCullough JP, Klein DN, Arnow B, Dunner DL, Gelenberg AJ, et al.A Comparison of Nefazodone, the cognitive behavioral-analysis system ofpsychotherapy, and their combination for the treatment of chronicdepression. New Eng J Med 2000; 342: 1462–70.

42 Kennedy SH, Konarski JZ, Segal ZV, Lau MA, Bieling PJ, McIntyre RS, et al.Differences in brain glucose metabolism between responders to CBT andvenlafaxine in a 16-week randomized controlled trial. Am J Psychiatry 2007;164: 778–88.

43 Markowitz JC, Kocsis JH, Bleiberg KL, Christos PJ, Sacks M. A comparativetrial of psychotherapy and pharmacotherapy for ‘pure’ dysthymic patients.J Affect Disord 2005; 89: 167–75.

44 Martin SD, Martin E, Rai SS, Richardson MA, Royall R. Brain blood flowchanges in depressed patients treated with interpersonal psychotherapy orvenlafaxine hydrochloride: preliminary findings. Arch Gen Psychiatry 2001;58: 641–8.

45 McKnight DL, Nelson-Gray RO, Barnhill J. Dexamethasone suppression testand response to cognitive therapy and antidepressant medication. BehaviorTherapy 1992; 23: 99–111.

46 McLean PD, Hakstian AR. Clinical depression: comparative efficacy ofoutpatient treatments. J Consult Clin Psychol 1979; 47: 818–36.

47 Menchetti M, Rucci P, Bortolotti B, Bombi A, Scocco P, Kraemer HC, et al.Moderators of remission with interpersonal counselling or drug treatmentin primary care patients with depression: randomised controlled trial.Br J Psychiatry 2014; 204: 144–50.

48 Miranda J, Chung JY, Green BL, Krupnick J, Siddique J, Revicki DA, et al.Treating depression in predominantly low-income young minority women:a randomized controlled trial. JAMA 2003; 290: 57–65.

49 Mohr DC, Boudewyn AC, Goodkin DE, Bostrom A, Epstein L. Comparativeoutcomes for individual cognitive-behavior therapy, supportive-expressivegroup psychotherapy, and sertraline for the treatment of depression inmultiple sclerosis. J Consult Clin Psychol 2001; 69: 942–9.

50 Moradveisi L, Huibers MJH, Renner F, Arasteh M, Arntz A. Behaviouralactivation v. antidepressant medication for treating depression in Iran:randomised trial. Br J Psychiatry 2013; 202: 204–11.

51 Murphy GE, Simons AD, Wetzel RD, Lustman PJ. Cognitive therapy andpharmacotherapy. Singly and together in the treatment of depression.Arch Gen Psychiatry 1984; 41: 33–41.

52 Mynors-Wallis LM, Gath DH, Day A, Baker F. Randomised controlled trial ofproblem solving treatment, antidepressant medication, and combinedtreatment for major depression in primary care. BMJ 2000; 320: 26–30.

53 Mynors-Wallis LM, Gath DH, Lloyd-Thomas AR, Tomlinson D. Randomisedcontrolled trial comparing problem solving treatment with amitriptyline andplacebo for major depression in primary care. BMJ 1995; 310: 441–5.

54 Parker G, Blanch B, Paterson A, Hadzi-Pavlovic D, Sheppard E,Manicavasagar V, et al. The superiority of antidepressant medication tocognitive behavior therapy in melancholic depressed patients: a 12-weeksingle-blind randomized study. Acta Psychiatr Scand 2013; 128: 271–81.

55 Quilty LC, McBride C, Bagby RM. Evidence for the cognitive mediationalmodel of cognitive behavioural therapy for depression. Psychol Med 2008;38: 1531–41.

56 Rush AJ, Beck AT, Kovacs M, Hollon S. Comparative efficacy of cognitivetherapy and pharmacotherapy in the treatment of depressed outpatients.Cogn Ther Res 1977; 1: 17–37.

57 Salminen JK, Karlsson H, Hietala J, Kajander J, Aalto S, Markkula J, et al.Short-term psychodynamic psychotherapy and fluoxetine in major depressivedisorder: a randomized comparative study. Psychother Psychosom 2008; 77:351–7.

58 Schulberg HC, Block MR, Madonia MJ, Scott CP, Rodriguez E, Imber SD, et al.Treating major depression in primary care practice. Eight-month clinicaloutcomes. Arch Gen Psychiatry 1996; 53: 913–9.

59 Scott AI, Freeman CP. Edinburgh primary care depression study: treatmentoutcome, patient satisfaction, and cost after 16 weeks. BMJ 1992; 304:883–7.

60 Shamsaei F, Rahimi A, Zarabian MK, Sedehi M. Efficacy of pharmacotherapyand cognitive therapy, alone and in combination in major depressivedisorder. Hong Kong J Psychiatry 2008; 18: 76.

61 Sharp DJ, Chew-Graham C, Tylee A, Lewis G, Howard L, Anderson I, et al.A pragmatic randomised controlled trial to compare antidepressants with acommunity-based psychosocial intervention for the treatment of womenwith postnatal depression: the RESPOND trial. Health Technol Assess 2010;14: iii–iv, ix–xi, 1–153.

62 Sloane RB, Staples FR, Schneider LS. Interpersonal therapy vs. nortriptylinefor depression in the elderly. In Clinical and Pharmacological Studies inPsychiatric Disorders (ed GD Burrows, TR Norman and L Dennerstein): 344–6.John Libbey, 1985.

63 Thompson LW, Coon DW, Gallagher-Thompson D, Sommer BR, Koin D.Comparison of desipramine and cognitive/behavioral therapy in thetreatment of elderly outpatients with mild-to-moderate depression.Am J Geriatr Psychiatry 2001; 9: 225–40.

64 Weissman MM, Prusoff BA, Dimascio A, Neu C, Goklaney M, Klerman GL. Theefficacy of drugs and psychotherapy in the treatment of acute depressiveepisodes. Am J Psychiatry 1979; 136: 555–8.

65 Williams JW, Barrett J, Oxman T, Frank E, Katon W, Sullivan M, et al.Treatment of dysthymia and minor depression in primary care: a randomizedcontrolled trial in older adults. JAMA 2000; 284: 1519–26.

66 Zu S, Xiang Y-T, Liu J, Zhang L, Wang G, Ma X, et al. A comparison ofcognitive-behavioral therapy, antidepressants, their combination andstandard treatment for Chinese patients with moderate-severe majordepressive disorders. J Affect Disord 2014; 152–4: 262–7.

67 Imel ZE, Malterer MB, McKay KM, Wampold BE. A meta-analysis ofpsychotherapy and medication in unipolar depression and dysthymia.J Affect Disord 2008; 110: 197–206.

68 Cuijpers P, Turner EH, Koole SL, van Dijke A, Smit F. What Is the threshold fora clinically relevant effect? The case of major depressive disorders. DepressAnxiety 2014; 31: 374–8.

69 Cipriani A, Furukawa TA, Salanti G, Geddes JR, Higgins JP, Churchill R, et al.Comparative efficacy and acceptability of 12 new-generationantidepressants: a multiple-treatments meta-analysis. Lancet 2009; 373:746–58.

70 Cuijpers P, van Straten A, Andersson G, van Oppen P. Psychotherapy fordepression in adults: a meta-analysis of comparative outcome studies.J Consult Clin Psychol 2008; 76: 909–22.

71 Okike K, Kocher MS, Mehlman CT, Bhandari M. Industry-sponsored research.Injury 2008; 39: 666–80.

72 Eisner M, Humphreys DK, Wilson P, Gardner F. Disclosure of financialconflicts of interests in interventions to improve child psychosocial health:a cross-sectional study. PLoS One 2015; 10: e0142803.

8

Data supplement to Cristea et al. Sponsorship bias in the comparative efficacy of psychotherapy and pharmacotherapy for adult depression: meta-analysis. Br J Psychiatry doi: 10.1192/bjp.bp.115.179275

Fig. DS1. Funnel plots. (a) Trials with industry funding; (b) Trials without industry funding (with imputed studies)

Fig. DS2. Funnel plots. (a) Trials with authors with financial COI; (b) Trials with no information on financial COI (with imputed studies)

Supplement DS1

Search string and list of included studies

A. Complete search string for PubMed

("psychotherapy"[MeSH Terms] OR "psychotherap*"[All Fields] OR cbt[All Fields] OR "cognitive behavior* therap*"[All Fields] OR "cognitive behavior* therap*"[All Fields] OR "behavior* therap*"[All Fields] OR "behavior* therap*"[All Fields] OR "cognition therap*"[All Fields] OR psychodynamic[All Fields] OR "psychoanalysis"[MeSH Terms] OR "psychoanalysis"[All Fields]) OR psychoanalytic*[All Fields] OR "counselling"[All Fields] OR "counseling"[MeSH Terms] OR "counseling"[All Fields]) OR "problem-solving"[All Fields] OR "problem solving"[All Fields] OR "mindfulness"[All Fields] OR (acceptance[All Fields] AND "commitment"[All Fields]) OR "assertiveness training"[All Fields] OR "behavior* activation"[All Fields] OR "cognitive therap*"[All Fields] OR "cognitive restructuring"[All Fields] OR "metacognitive therap*"[All Fields] OR "solution-focused therap*"[All Fields] OR "self-control therap*"[All Fields] OR "self control therap*"[All Fields] OR "self-control training"[All Fields] OR "self-control training"[All Fields]) AND ("depressive disorder"[MeSH Terms] OR ("depressive"[All Fields] AND "disorder"[All Fields]) OR "depressive disorder"[All Fields] OR "depression"[All Fields] OR "depression"[MeSH Terms]) OR depressive[All Fields] OR "major depression"[All Fields] OR "major depressive disorder"[All Fields] OR "dysthymic disorder"[MeSH Terms] OR ("dysthymic"[All Fields] AND "disorder"[All Fields]) OR "dysthymic disorder"[All Fields] OR "dysthymia"[All Fields] OR dysthymic[All Fields] OR "mood disorder"[All Fields] OR "affective disorder"[All Fields]) Filters: Randomized Controlled Trial

B. List of included studies

1 Barber JP, Barrett MS, Gallop R, Rynn MA, Rickels K. Short-term dynamic psychotherapy versus pharmacotherapy for major depressive disorder: a randomized, placebo-controlled trial. J Clin Psychiatry 2012; 73: 66–73. 2 Barrett JE, Williams JW, Oxman TE, Frank E, Katon W, Sullivan M, et al. Treatment of dysthymia and minor depression in primary care: a randomized trial in patients aged 18 to 59 years. J Fam Pract 2001; 50: 405–12. 3 Bedi N, Chilvers C, Churchill R, Dewey M, Duggan C, Fielding K, et al. Assessing effectiveness of treatment of depression in primary care. Partially randomised preference trial. Br J Psychiatry 2000; 177: 312–8. 4 Blackburn IM, Moore RG. Controlled acute and follow-up trial of cognitive therapy and pharmacotherapy in out-patients with recurrent depression. Br J Psychiatry 1997; 171: 328–34. 5 Blom MBJ, Jonker K, Dusseldorp E, Spinhoven P, Hoencamp E, Haffmans J, et al. Combination treatment for acute depression is superior only when psychotherapy is added to medication. Psychother Psychosom 2007; 76: 289–97. 6 Browne G, Steiner M, Roberts J, Gafni A, Byrne C, Dunn E, et al. Sertraline and/or interpersonal psychotherapy for patients with dysthymic disorder in primary care: 6-month comparison with longitudinal 2-year follow-up of effectiveness and costs. J Affect Disord 2002; 68: 317–30. 7 David D, Szentagotai A, Lupu V, Cosman D. Rational emotive behavior therapy, cognitive therapy, and medication in the treatment of major depressive disorder: a randomized clinical trial, posttreatment outcomes, and six-month follow-up. J Clin Psychol 2008; 64: 728–46.

8 Dekker JJM, Koelen JA, Van HL, Schoevers RA, Peen J, Hendriksen M, et al. Speed of action: the relative efficacy of short psychodynamic supportive psychotherapy and pharmacotherapy in the first 8 weeks of a treatment algorithm for depression. J Affect Disord 2008; 109: 183–8. 9 DeRubeis RJ, Hollon SD, Amsterdam JD, Shelton RC, Young PR, Salomon RM, et al. Cognitive therapy vs medications in the treatment of moderate to severe depression. Arch Gen Psychiatry 2005; 62: 409–16. 10 Dimidjian S, Hollon SD, Dobson KS, Schmaling KB, Kohlenberg RJ, Addis ME, et al. Randomized trial of behavioral activation, cognitive therapy, and antidepressant medication in the acute treatment of adults with major depression. J Consult Clin Psychol 2006; 74: 658–70. 11 Dunlop BW, Kelley ME, Mletzko TC, Velasquez CM, Craighead WE, Mayberg HS. Depression beliefs, treatment preference, and outcomes in a randomized trial for major depressive disorder. J Psychiatr Res 2012; 46: 375–81. 12 Dunner DL, Schmaling KB, Hendrickson H, Becker J, Lehman A, Bea C. Cognitive therapy versus fluoxetine in the treatment of dysthymic disorder. Depression 1996; 4: 34–41. 13 Elkin I, Shea MT, Watkins JT, Imber SD, Sotsky SM, Collins JF, et al. National Institute of Mental Health Treatment of Depression Collaborative Research Program. General effectiveness of treatments. Arch Gen Psychiatry 1989; 46: 971–82; discussion 983. 14 Faramarzi M, Alipor A, Esmaelzadeh S, Kheirkhah F, Poladi K, Pash H. Treatment of depression and anxiety in infertile women: cognitive behavioral therapy versus fluoxetine. J Affect Disord 2008; 108: 159–64. 15 Finkenzeller DW, Zobel I, Rietz S, Schramm E, Berger M. Interpersonelle Psychotherapie und Pharmakotherapie bei Post-Stroke-Depression. Nervenarzt 2009; 80: 805–12. 16 Frank E, Cassano GB, Rucci P, Thompson WK, Kraemer HC, Fagiolini A, et al. Predictors and moderators of time to remission of major depression with interpersonal psychotherapy and SSRI pharmacotherapy. Psychol Med 2011; 41: 151–62. 17 Hegerl U, Hautzinger M, Mergl R, Kohnen R, Schütze M, Scheunemann W, et al. Effects of pharmacotherapy and psychotherapy in depressed primary-care patients: a randomized, controlled trial including a patients’ choice arm. Int J Neuropsychopharmacol 2010; 13: 31–44. 18 Hollon SD, DeRubeis RJ, Evans MD, Wiemer MJ, Garvey MJ, Grove WM, et al. Cognitive therapy and pharmacotherapy for depression. Singly and in combination. Arch Gen Psychiatry 1992; 49: 774–81. 19 Jarrett RB, Schaffer M, McIntire D, Witt-Browder A, Kraft D, Risser RC. Treatment of atypical depression with cognitive therapy or phenelzine: a double-blind, placebo-controlled trial. Arch Gen Psychiatry 1999; 56: 431–7. 20 Keller MB, McCullough JP, Klein DN, Arnow B, Dunner DL, Gelenberg AJ, et al. A Comparison of Nefazodone, the Cognitive Behavioral-Analysis System of Psychotherapy, and Their Combination for the Treatment of Chronic Depression. New England Journal of Medicine 2000; 342: 1462–70. 21 Kennedy SH, Konarski JZ, Segal ZV, Lau MA, Bieling PJ, McIntyre RS, et al. Differences in brain glucose metabolism between responders to CBT and venlafaxine in a 16-week randomized controlled trial. Am J Psychiatry 2007; 164: 778–88. 22 Markowitz JC, Kocsis JH, Bleiberg KL, Christos PJ, Sacks M. A comparative trial of psychotherapy and pharmacotherapy for ‘pure’ dysthymic patients. J Affect Disord 2005; 89: 167–75. 23 Martin SD, Martin E, Rai SS, Richardson MA, Royall R. Brain blood flow changes in depressed patients treated with interpersonal psychotherapy or venlafaxine hydrochloride: preliminary findings. Arch Gen Psychiatry 2001; 58: 641–8. 24 McKnight DL, Nelson-Gray RO, Barnhill J. Dexamethasone suppression test and response to cognitive therapy and antidepressant medication. Behavior Therapy 1992; 23: 99–111. 25 McLean PD, Hakstian AR. Clinical depression: comparative efficacy of outpatient treatments. J Consult Clin Psychol 1979; 47: 818–36.

26 Menchetti M, Rucci P, Bortolotti B, Bombi A, Scocco P, Kraemer HC, et al. Moderators of remission with interpersonal counselling or drug treatment in primary care patients with depression: randomised controlled trial. Br J Psychiatry 2014; 204: 144–50. 27 Miranda J, Chung JY, Green BL, Krupnick J, Siddique J, Revicki DA, et al. Treating depression in predominantly low-income young minority women: a randomized controlled trial. JAMA 2003; 290: 57–65. 28 Mohr DC, Boudewyn AC, Goodkin DE, Bostrom A, Epstein L. Comparative outcomes for individual cognitive-behavior therapy, supportive-expressive group psychotherapy, and sertraline for the treatment of depression in multiple sclerosis. J Consult Clin Psychol 2001; 69: 942–9. 29 Moradveisi L, Huibers MJH, Renner F, Arasteh M, Arntz A. Behavioural activation v. antidepressant medication for treating depression in Iran: randomised trial. Br J Psychiatry 2013; 202: 204–11. 30 Murphy GE, Simons AD, Wetzel RD, Lustman PJ. Cognitive therapy and pharmacotherapy. Singly and together in the treatment of depression. Arch Gen Psychiatry 1984; 41: 33–41. 31 Mynors-Wallis LM, Gath DH, Day A, Baker F. Randomised controlled trial of problem solving treatment, antidepressant medication, and combined treatment for major depression in primary care. BMJ 2000; 320: 26–30. 32 Mynors-Wallis LM, Gath DH, Lloyd-Thomas AR, Tomlinson D. Randomised controlled trial comparing problem solving treatment with amitriptyline and placebo for major depression in primary care. BMJ 1995; 310: 441–5. 33 Parker G, Blanch B, Paterson A, Hadzi-Pavlovic D, Sheppard E, Manicavasagar V, et al. The superiority of antidepressant medication to cognitive behavior therapy in melancholic depressed patients: a 12-week single-blind randomized study. Acta Psychiatr Scand 2013; 128: 271–81. 34 Quilty LC, McBride C, Bagby RM. Evidence for the cognitive mediational model of cognitive behavioural therapy for depression. Psychol Med 2008; 38: 1531–41. 35 Rush AJ, Beck AT, Kovacs M, Hollon S. Comparative efficacy of cognitive therapy and pharmacotherapy in the treatment of depressed outpatients. Cogn Ther Res 1977; 1: 17–37. 36 Salminen JK, Karlsson H, Hietala J, Kajander J, Aalto S, Markkula J, et al. Short-term psychodynamic psychotherapy and fluoxetine in major depressive disorder: a randomized comparative study. Psychother Psychosom 2008; 77: 351–7. 37 Schulberg HC, Block MR, Madonia MJ, Scott CP, Rodriguez E, Imber SD, et al. Treating major depression in primary care practice. Eight-month clinical outcomes. Arch Gen Psychiatry 1996; 53: 913–9. 38 Scott AI, Freeman CP. Edinburgh primary care depression study: treatment outcome, patient satisfaction, and cost after 16 weeks. BMJ 1992; 304: 883–7. 39 Shamsaei F, Rahimi A, Zarabian MK, Sedehi M. Efficacy of Pharmacotherapy and Cognitive Therapy, Alone and in Combination in Major Depressive Disorder. Hong Kong Journal of Psychiatry 2008; 18: 76. 40 Sharp DJ, Chew-Graham C, Tylee A, Lewis G, Howard L, Anderson I, et al. A pragmatic randomised controlled trial to compare antidepressants with a community-based psychosocial intervention for the treatment of women with postnatal depression: the RESPOND trial. Health Technol Assess 2010; 14: iii – iv, ix – xi, 1–153. 41 Sloane RB, Staples FR, Schneider LS. Interpersonal therapy vs. nortriptyline for depression in the elderly. Clinical and pharmacological studies in psychiatric disorders Biological psychiatry - new prospects; 1985: 344–6. 42 Thompson LW, Coon DW, Gallagher-Thompson D, Sommer BR, Koin D. Comparison of desipramine and cognitive/behavioral therapy in the treatment of elderly outpatients with mild-to-moderate depression. Am J Geriatr Psychiatry 2001; 9: 225–40. 43 Weissman MM, Prusoff BA, Dimascio A, Neu C, Goklaney M, Klerman GL. The efficacy of drugs and psychotherapy in the treatment of acute depressive episodes. Am J Psychiatry 1979; 136: 555–8.

44 Williams JW, Barrett J, Oxman T, Frank E, Katon W, Sullivan M, et al. Treatment of dysthymia and minor depression in primary care: A randomized controlled trial in older adults. JAMA 2000; 284: 1519–26. 45 Zu S, Xiang Y-T, Liu J, Zhang L, Wang G, Ma X, et al. A comparison of cognitive-behavioral therapy, antidepressants, their combination and standard treatment for Chinese patients with moderate-severe major depressive disorders. J Affect Disord 2014; 152-154: 262–7.

Table DS1 Selected characteristics of studies directly comparing psychotherapy and pharmacotherapy for adult depression

Study Recr Target grp

Diagn Psych Npsy Nsess Deliv Med Npha Study financing COI COI source % AU COI

RoBa)

SG AG BA ITT Bas Sev (HRSD)

Prov

Barber, 2011 Comm Adults MDD DYN 51 20 Ind Mix/Oth 55 GOV+PHA(FM) Y Orig.art 60% + ? + + 19.43 US

Barrett, 2001 Clin Adults Mood PST 80 6 Ind SSRI 80 OTH+PHA (FM)

Y Frank, 2011

11% + ? + − 14.15 US

Bedi, 2000 Clin Adults MDD Couns 39 NR Ind Mix/Oth 44 GOV NR ? + SR − - EU

Blackburn,1997 Clin Adults MDD CBT 24 16 Ind Mix/Oth 43 GOV NR ? ? + − 20.05 EU

Blom, 2007 Clin Adults MDD IPT 34 12 Ind SNRI 30 GOV+PHA NR ? ? + − 21.07 EU Browne, 2002 Comm Adults DYS IPT 122 10 Ind SSRI 117 GOV+PHA NR + + + − - CA David, 2008 Comm Adults MDD CBT 56 20 Ind SSRI 57 GOV+OTH NR ? ? + + 22.46 EU REBT 57 20 Ind

Dekker, 2008 Clin Adults MDD DYN 59 16 Ind Mix/Oth 44 PHA NR ? ? + − 20.14 EU

DeRubeis, 2005 Comm Adults MDD CBT 60 14 Ind SSRI 120 GOV+PHA(FM) NR ? ? + + 21.52 US

Dimidjian, 2006 Comm Adults MDD CBT 35 16 Ind SSRI 49 GOV+PHA(FM) Y Orig.art 8% + − + + 20.62 US

BAT 29 16 Ind

Dunlop, 2012 Comm Adult MDD CBT 41 16 Ind SSRI 39 PHA Y Orig.art 33% ? + + − 19.70 US

Dunner, 1996 NR Adults DYS CBT 9 16 Ind SSRI 11 NR NR ? ? + − 15.99 US

Elkin, 1989 Clin Adults MDD IPT 61 16 Ind TCA 57 GOV+OTH NR + ? + − 19.57 US

CBT 59 16 Ind

Faramarzi, 2008 Other MDD CBT 29 10 Grp SSRI 30 OTH NR ? ? SR − - Iran

Finkenzeller, 2009 Other Stroke MDD IPT 23 12 Grp SSRI 24 PHA NR + ? + + 21.03 EU Frank, 2011 Clin Adults MDD IPT 160 12 Ind SSRI 158 GOV+PHA Y Orig.art 40% − − + + 20 US

Hegerl, 2010 Clin Adults Mood CBT 52 10 Grp SSRI 76 GOV Y Orig.art 22% + + + + 16.58 EU

Hollon, 1992 Clin Adults MDD CBT 25 20 Ind TCA 57 GOV+OTH NR ? ? + + 23.89 US

Jarrett, 1999 Comm Adults MDD CBT 36 20 Ind MAOI 36 GOV+PHA(FM) NR ? + + + - US

Keller, 2000 Clin Adults Mood CBASP 226 18 Ind SNRI 220 PHA Y Orig.art 92% + + + 0 - US

Kennedy, 2007 Clin Adults MDD CBT 12 16 Ind SNRI 12 GOV+PHA Y Orig. art 42% ? ? ? − 21.90 CA

Markowitz, 2005 Comm Adults DYS IPT 23 17 Ind SSRI 24 GOV+PHA Y Indep artb) 20% + − + + - US

Couns 26 17 Ind

Martin, 2001 Comm Adults MDD IPT 13 16 Ind SNRI 15 PHA Y Indep art b) 20% − − − + 22.50 EU

McKnight, 1992 Comm Adults MDD CBT 12 8 Ind TCA 11 NR NR ? ? ? − - US

McLean, 1979 Comm Adults MDD CBT 42 10 Ind TCA 49 GOV NR ? ? SR − - CA

DYN 44 10 Ind

Menchetti, 2014 Clin Adults MDD IPT 143 6 Ind SSRI 144 GOV Y Frank, 2011 & Indep art b)

43% + + ? + - EU

Miranda, 2003 Other Other MDD CBT 90 8 Ind/Grp Mix/Oth 88 GOV+PHA(FM) NR + + + + 17.14 US

Mohr, 2001 Other MS patients

MDD CBT 20 16 Ind SSRI 15 GOV+OTH NR − − − + 20.7 US

SEG 19 16 Grp

Moradvesi, 2013 Comm Adults MDD BAT 50 16 Ind SSRI 50 OTH NR + + + + 21.37 Iran

Murphy, 1984 Clin Adults MDD CBT 24 20 Ind TCA 24 GOV+PHA (FM)

NR − − − − 19.88 US

Mynors-Wallis, 1995 Clin Adults MDD PST 29 6 Ind TCA 27 OTH+PHA(FM) NR ? + + − 19.26 EU

Mynors-Wallis, 2000 Clin Adults MDD PST gp 39 6 Ind SSRI 36 GOV NR + + + + 20.41 EU

PST n 41 6 Ind

Quilty, 2008 Comm Adults MDD CBT 45 16 Ind Mix/Oth 30 OTH NR ? ? ? − - CA

IPT 46 16 Ind

Parker, 2013 Comm Adults Mood CBT 11 22 Ind Mix/Oth 10 GOV+PHA(FM) Y Orig.art 9% 17.57 AU

Rush, 1977 Clin Adults MDD CBT 19 20 Ind TCA 22 GOV+OTH NR ? ? + − 30.15 US

Salminen, 2008 Clin Adults MDD DYN 26 16 Ind SSRI 25 GOV+OTH Y Indep art b) 38% ? ? − + 18.59 EU

Schulberg, 1996 Adults Clin MDD IPT 93 16 Ind TCA 91 NR NR ? ? + + 22.99 US

Scott, 1992 Adults Clin MDD CBT 29 16 Ind TCA 26 GOV NR ? + + − 17.37 EU

Couns 29 16 Ind TCA

Shamsaei, 2008 Clin Adults MDD CBT 40 8 Ind SSRI 40 OTH NR ? + SR ? - Iran

Sharp, 2010 Other PPD MDD Couns 112 6 Ind Mix/Oth 106 GOV Y Orig.art 42% + + + + - EU

Sloane, 1985 Other Elderly MDD IPT 19 6 Ind TCA 10 NR NR − − + − 23.2 US

Thompson, 2001 Comm Elderly MDD CBT 36 18 Ind TCA 33 GOV+PHA(FM) NR ? ? ? − 18.85 US

Weissman, 1979 Clin Adults MDD IPT 23 16 Ind TCA 20 GOV NR ? ? + − - US

Williams, 2000 Comm Elderly Mood PST 113 6 Ind SSRI 106 OTH+PHA (FM)

Y Orig.art 13% + + + + 13.55 US

Zu, 2014 Clin Adults MDD CBT 30 20 Ind SSRI 60 GOV NR + ? + − 22.03 CH

a) RoB: risk of bias according to the Cochrane Collaboration tool. SG, sequence generation; AC, allocation concealment; BA; blinding of assessors; ITT, intent-to-treat analysis to handle missing data. Ratings of “+” indicate the study has a low RoB on that criteria; ratings of “-“ indicate high RoB; “?” uncertain RoB; SR, only self-report measures

b) Independent articles in order of appearance: Epstein J, Pan H, Kocsis JH, et al. Lack of Ventral Striatal Response to Positive Stimuli in Depressed Versus Normal Subjects. AJP 2006;163:1784–90. doi:10.1176/ajp.2006.163.10.1784; Fleischhacker WW, Eerdekens M, Karcher K, et al. Treatment of schizophrenia with long-acting injectable risperidone: a 12-month open-label trial of the first long-acting second-generation antipsychotic. J Clin Psychiatry 2003;64:1250–7; 3.Hirvonen J, Karlsson H, Kajander J, et al. Decreased brain serotonin 5-HT1A receptor availability in medication-naive patients with major depressive disorder: an in-vivo imaging study using PET and [carbonyl-11C]WAY-100635. Int J Neuropsychopharmacol 2008;11:465–76. doi:10.1017/S1461145707008140; Rucci P, Frank E, Scocco P, et al. Treatment-emergent suicidal ideation

during 4 months of acute management of unipolar major depression with SSRI pharmacotherapy or interpersonal psychotherapy in a randomized clinical trial. Depress Anxiety 2011;28:303–9. doi:10.1002/da.20758

Abbreviations: BAT: behavioral activation therapy; Bas Sev; Baseline Severity (evaluated on the HRSD-17); CA: Canada; CH, China; CBASP: cognitive behavioral analysis system of psychotherapy; CBT: cognitive behavior therapy; Clin: recruitment from clinical samples only; Comm: (part of the) sample recruited from the community; Couns: non-directive supportive counseling; Deliv: delivery; Diagn: diagnosis; DYN: psychodynamic therapy; DYS: dysthymic disorder; EU: Europe; FM: free medication only; GOV: government funding; Grp: group format; Ind: individual format; Indep art: independent article; IPT: interpersonal psychotherapy; MAOI: Monoamine oxidase inhibitor; MDD: major depressive disorder; Med: medication; Mix/oth: other antidepressant, mix of antidepressants or protocolized treatment with antidepressants; Mood: mixed/other mood disorder (e.g., minor depression, chronic depression, melancholic depression); MS: multiple sclerosis; Nsess: number of sessions; Npha: number of patients in the pharmacotherapy conditions; Npsy: number of patients in the psychotherapy condition; NR: not reported; OTH: funding by other institution; Orig.art: original article; Psych: psychotherapy; PHA: funding by a pharmacological company; PPD: post-partum depression; Prov: provenience; PST gp: PST by a general practitioner; PST n: PST by a nurse; PST: problem-solving therapy; REBT: rational emotive behavior therapy; Recr: recuitment; SNRI: Serotonin–norepinephrine reuptake inhibitor; Specific, SSRI: Selective serotonin reuptake inhibitor; SEG: supportive-expressive group therapy; TCA: Tricyclic antidepressant; US: United States; Y, yes.

10.1192/bjp.bp.115.179275 published online November 3, 2016 Access the most recent version at DOI: BJP 

Ioana A. Cristea, Claudio Gentili, Pietro Pietrini and Pim Cuijpersand pharmacotherapy for adult depression: meta-analysisSponsorship bias in the comparative efficacy of psychotherapy

MaterialSupplementary

http://bjp.rcpsych.org/content/suppl/2016/10/21/bjp.bp.115.179275.DC1.htmlSupplementary material can be found at:

Referenceshttp://bjp.rcpsych.org/content/early/2016/10/20/bjp.bp.115.179275#BIBLThis article cites 0 articles, 0 of which you can access for free at:

permissionsReprints/

permissions@rcpsych.ac.ukwrite to To obtain reprints or permission to reproduce material from this paper, please

P<P Published online 2016-11-03T00:05:10-07:00 in advance of the print journal.

to this article atYou can respond /letters/submit/bjprcpsych;bjp.bp.115.179275v1

from Downloaded

The Royal College of PsychiatristsPublished by on November 3, 2016http://bjp.rcpsych.org/

digital object identifier (DOIs) and date of initial publication. theindexed by PubMed from initial publication. Citations to Advance online articles must include

final publication). Advance online articles are citable and establish publication priority; they areappeared in the paper journal (edited, typeset versions may be posted when available prior to Advance online articles have been peer reviewed and accepted for publication but have not yet

http://bjp.rcpsych.org/site/subscriptions/ go to: The British Journal of PsychiatryTo subscribe to